The purpose of this project is to search for susceptibility genes for autism using intermediate phenotypes. We will test 8-10 large extended Utah pedigrees, and 50 smaller multiplex pedigrees (2-4 affected cases). Seven large pedigrees and 57 smaller multiplex pedigrees have already been identified, and we have preliminary evidence that extensive phenotyping in these pedigrees will be feasible. Pedigrees will be typed with microsatellite markers and the Affymetrix 10k SNP chip. The phenotypes we have chosen to measure on these pedigrees show evidence of heritable variation. New pedigrees will be identified through the ascertainment of 800 trio families (86 trios are already complete and 214 more waiting to be scheduled at Utah, and 100 will come from our collaborating site in Colorado). The Utah Population DataBase, used to identify the first extended pedigrees, will be used to find new extended/multiplex pedigrees. Blood for DNA and selected phenotypes is collected on the trio families as they enter the study. This up-front DNA and phenotype collection provides an immediate resource not only for pedigree identification but also for later follow-up of positive linkage results and for the study of the best positional candidate genes found through the linkage analyses. The measurement of intermediate phenotypes on the trios as well as the pedigrees will allow us to stratify our sample to a degree that has not before been possible in large-scale genetic studies of autism. University of Utah molecular genetics labs will give us world-class expertise in genotyping, fine mapping, candidate gene selection, sequencing, and interpretation of results. Our specific aims are: 1) Ascertain 550 more trios for a total of 800 to find new multiplex/extended pedigrees, to use for linkage follow-up, and to study 2 positional candidate genes once genome scanning (Aim 2) is complete. 2) Perform additional specific phenotyping in 8-10 extended autism families and 50 smaller multiplex families. Test intermediate phenotypes for heritable variation. Use a whole genome scan to search for linkage using intermediate phenotypes. Fine map regions of interest. Study the two best positional candidate genes using pedigrees and trios.